NotebookLM just got a major chat upgrade with 8x more context, better memory, and smarter responses (www.techradar.com)

Google has upgraded NotebookLM to use Gemini’s full one‑million‑token context window across all plans, an ~8× jump in short‑term context that lets the assistant ingest and reason over entire books, long meeting transcripts, and sprawling project archives in a single chat. Google also sextupled the service’s memory store and added persistent, private conversation history and new "Goals" controls that let you set the assistant’s role and behavior (e.g., rigorous thesis advisor, Socratic editor, or game‑master). Google reports internal tests showing roughly a 50% uplift in user satisfaction for long, multi‑source responses. Technically, the update means NotebookLM can maintain far more multi‑document state and follow dialogue threads much deeper (Google says it can track context six times farther in conversations), reducing the need to copy‑paste excerpts and enabling richer cross‑referencing, citation support, and project continuity. The changes position NotebookLM closer to purpose‑built workspace agents (Notion AI, ChatGPT memory, Perplexity) while leveraging Google’s app/web integrations. For researchers and teams, this improves large‑scale document analysis, continuity across interrupted work sessions, and customizable assistant behavior—while keeping chat history private unless explicitly shared.